[Gcl-devel] FFI sized integer data types

[Gabriel Dos Reis]

Hi Camm,

  I'd like to ask for a feature request:  that GCL makes official
support for the following ISO C sized integer datatypes (and their or
course, their corresponding unsigned counterparts)

    int8_t                      // 8-bit integer
    int16_t                     // 16-bit integer
    int32_t                     // 32-bit integer
    int64_t                     // 64-bit integer
    intptr_t                    // integer wide enough to hold a pointer


These are essential for systems such as OpenAxiom that needs to
interface with the hosting operating system services.

At the moment, I use the correspondance fixnum <=> intptr_t, but I'm
doinb this only because I've looked at the implementation details and
know things would break if that weren't true.  However, for the others
there does not seem to be any correspoding GCL-level FFI type.

FWIW, CLisp and ECL have support for those types;  SBCL has support
for almost all of them.

Bes,

-- Gaby